Published on Web 09/25/2003
Distinguishing Rate-Limiting Electron versus H-Atom Transfers in Cu2(O2)-Mediated Oxidative N-Dealkylations: Application of Inter- versus Intramolecular Kinetic Isotope Effects Jason Shearer, Christiana Xin Zhang, Lanying Q. Hatcher, and Kenneth D. Karlin* Department of Chemistry, Johns Hopkins UniVersity, 3400 North Charles Street, Baltimore, Maryland 21218 Received May 4, 2003; E-mail:
[email protected]
Hydroxylation reactions performed by Cu(I)-dioxygen adducts are biologically important; yet the diverse nature of active site structures and substrate types leaves many mechanistic questions unresolved.1-3 For example, tyrosinase o-phenol hydoxylations (proceeding from a CuII2-µ-η2-η2-peroxo species) appear to occur via an electrophilic mechanism.1,4 However, recent model studies by Tolman and Itoh suggest that CuII2-peroxo/CuIII2-bis-µ-oxo complexes are capable of oxidizing substrates through rate-limiting hydrogen atom transfer (HAT) pathways.2a,5,6 Studies on dopamineβ-hydroxylase (DβH) and for peptide oxidative N-dealkyation by peptidylglycine R-hydroxylating monooxygenase (PHM) previously implicated Cu-hydroperoxo or Cu-superoxo species facilitating observed HAT reactions; however, recent insights suggest that alternative copper-dioxygen derived active species need to be considered.3 To better understand how CuII-peroxo species oxidize substrates, we recently reported on the preparation of a series of CuI complexes, [CuI(MePY2)R′]+, where Cu is contained within bis[2-(2-(4-R′pyridyl)ethyl]methylamine tridentate ligands (MePY2R′, R′ ) H, MeO, Me2N; Scheme 1).7 These complexes readily react with dioxygen, forming the corresponding Cu2II-O2 adducts [(Cu2II(MePY2)R′)2(O2)]2+ (1R′, R′ ) H, MeO, Me2N), where the Cu2IIperoxo complex is in equilibrium with the corresponding Cu2IIIbis-µ-oxo adduct.7-9 Also, 1R′ readily oxidize substrates such as tetrahydrofuran (THF), alcohols, and N,N-dimethylaniline (DMA).9 para-Substituted DMAs (R-DMAs) have been used as mechanistic probes, distinguishing between rate-limiting HAT or electrontransfer (ET) pathways, for example in cytochrome P450 (P450) chemistry (Scheme 2).10 Here, we wish to communicate that the use of R-DMAs has yielded rich new insight into the nature of oxidations induced by Cu(I)-dioxygen adducts. In fact, oxidations by 1R′ can occur through both a rate-limiting ET or a HAT pathway, as has been suggested for high valent Fe-oxo porphyrinates.10,11 Dichloromethane solutions of dioxygen adducts 1R′ under argon (with excess O2 removed) at -80 °C readily react with R-DMA (R ) MeO, Me, H, CN), affording the corresponding parasubstituted N-methylaniline (R-MA) and formaldehyde in good yields.7,12,13 With N,N-dibenzylaniline as substrate, isolation of the benzaldehyde product from O2 versus 18O2 reactions14 suggests a “rebound” type mechanism analogous to P450 chemistry. This indicates an overall C-H bond homolysis proceeding through either an ET followed by a proton transfer (PT), or a HAT pathway (Scheme 2a and b, respectively). Because the oxidative N-dealkylation yields of R-MA closely compare for a given 1R′ (Table S1),13 we can determine the relative rates of these reactions using competition studies and measured R-MA yields. Oxidative competition reactions induced by 1H run in a 1:1 mixture of R-DMA:H-DMA demonstrate a strong R-group dependence on the relative rates (krel). As R is made more electron-donating, krel increases (Table 1). A linear free-energy correlation gives a large negative F value (F ) -2.1, r2 ) 0.99).14 12670
9
J. AM. CHEM. SOC. 2003, 125, 12670-12671
Scheme 1
Scheme 2
Table 1. krel: R-DMA Competition Studies (CH2Cl2, -80 °C)16
MeO-DMA Me-DMA H-DMA CN-DMA F (r2)
1H
1Me2N
1MeO
σ+
11.4 (3) 2.5 (1) 1.0 0.02 (1) -2.1 (0.99)
2.3 (2) 1.2 (1) 1.0 0.53 (3) -0.47 (0.98)
12.4 (2) 2.2 (1) 1.0 0.48 (1) -0.99 (0.92)
-0.65 -0.26 0.00 0.67
Table 2. KIEintra versus KIEinter for the N-Dealkylation of R-DMA in CH2Cl2 at - 80 °C17
MeO-DMA Me-DMA H-DMA CN-DMA
a
1H a
1Me2N a
1MeO a
E1/2 b
4.7 (8)/ 1.7 (6) 4.6 (6)/ 2.3 (3) 4.1 (6)/ 2.4 (6) 2.6 (8)/ 2.1 (8)
7.3 (4)/ 2.7 (2) 5.8 (8)/ 3.8 (2) 12.0 (9)/ 11.4 (15) 14.9 (7)/ 15.0 (4)
7.5 (2)/ 2.3 (4) 5.0 (7)/ 3.0 (2) 6.1 (7)/ 2.7 (6) 13.9 (11)/ 13.1 (19)
0.53 0.72 0.92 1.21
KIEintra/KIEinter. b CH2Cl2 at room temperature (V vs SCE).10b
Scheme 3
Such a situation is suggestive of a rate-limiting ET process,15 followed by a PT from the DMA radical cation to the Cu-oxo core. This rate-limiting ET mechanism is also supported by the intraand intermolecular deuterium kinetic isotope effect profiles (KIEintra and KIEinter, see Table 2 and Scheme 3).10a,b,13 In the case of the intramolecular N-dealkylation reactions, the KIEinter profile for 1H shows a sharp increase as σ+ (and E1/2) for R-DMAs become more 10.1021/ja0359409 CCC: $25.00 © 2003 American Chemical Society
COMMUNICATIONS
negative, eventually reaching an asymptote (Table 2, Figure S4). Better H versus D differentiation occurs because the proton-transfer step becomes slower with DMA radical-cation stabilization by the electron-donating group. This translates into a larger KIEintra. In the case of the intermolecular reaction, there is a negligible difference in the isotope effect (KIEinter) as σ+ becomes more negative (Table 2), indicating that the ET event is mostly irreversible. If there was a reversible preequilibrium ET followed by a rate-limiting PT (peET/PT), one would expect to observe a KIEinter profile that increases as σ+ becomes more positive.10g The flat KIEinter profile indicates that the PT step has little influence on the overall oxidation of R-DMA by 1H. In other words, the product is determined by the (mostly irreversible) ET in the intermolecular reaction, and not the PT step. A rate-limiting ET is also supported by comparison of the absolute values obtained for KIEintra versus KIEinter (see Scheme 3). This is a powerful mechanistic probe for distinguishing between an ET or a HAT process.10e For a HAT mechanism, the KIEintra should be nearly identical to the KIEinter.10e This is because the rate of HAT versus deuterium atom transfer will be proportional to the C-H versus C-D bond dissociation enthalpies (BDEs). The difference in BDEs should be approximately the same in the intraversus the intermolecular reaction. In the case of the ET process, the expectation is that KIEinter < KIEintra.10e This is because in the intermolecular reaction the product will be determined by the ET event, while in the intramolecular reaction the PT event can potentially determine the product. For 1H, the values obtained for KIEinter are all less than those obtained for KIEintra, which fully supports a rate-limiting ET pathway for the oxidative N-dealkylation of R-DMA (Table 2, Scheme 2a). Also, both KIEintra and KIEinter values are relatively small in magnitude, in line with a rate-limiting ET mechanism.17 The situation is different in the case of 1Me2N. Competition reactions do not show a strong R-group dependence, with krel increasing only slightly as R is made more electron donating, Table 1. This is reflected in the linear free-energy correlation13 which yielded a F value consistent with either ET or HAT (F ) -0.49, r2 ) 0.98).15 The KIE profiles are largely inconclusive (Table 2), showing no distinct pattern for either HAT or ET.13 In the case of both KIEinter and KIEintra, what is observed is a general increase in KIE as σ+ becomes more positive. Furthermore, the KIEs become large in magnitude, consistent with a rate-limiting C-H bond cleavage. This could occur through a switch in mechanism from rate-limiting ET, to either a HAT or a peET/PT.10g A comparison of the magnitudes of the KIEinter versus KIEintra using the criterion mentioned above sheds further light on our results. For R ) MeO and Me, the data suggest that 1Me2N oxidizes R-DMA through a rate-limiting ET mechanism (KIEinter < KIEintra), while for the less reducing R-DMAs (H and CN), oxidation appears to occur through a rate-limiting HAT (KIEinter ≈ KIEintra). This is strong evidence in favor of a HAT mechanism. In addition, we favor the HAT over a peET/PT mechanism, as follows: In the case of 1H, we established rate-limiting ET (vide supra). However, 1Me2N is a weaker one-electron oxidant,18 and the µ-oxo groups in its Cu2O2 moiety should be more basic (as it possesses the stronger donor ligand MePY2Me2N).7 Thus, one would expect slower electron transfer and faster proton transfer in reactions of R-DMAs with 1Me2N relative to 1H; that is, ET would still be rate-limiting. Yet, the KIE values and criteria indicate this is not the case. Thus, peET/ PT is unlikely, and we conclude that HAT is operative for H- and CN-DMA in oxidations with 1Me2N. Other precedent comes from (a) that P450 may operate in a similar manner (ET for easily oxidized substrates and HAT for others),11 while (b) studies per-
formed by Tolman and Itoh suggest that Cu2O2 complexes are capable of performing HAT reactions from alkyl- and benzylamines.5,6 It therefore appears reasonable that as R-DMAs become harder to oxidize, there is a shift in mechanism for oxidative N-dealkylation by copper-dioxygen adduct 1Me2N from ET to HAT. By similar criteria, a changeover in mechanism is also suggested for 1MeO (data in Table 2) where the less easily oxidized CN-DMA reacts via a rate-limiting HAT pathway and the other substrates (R ) H, Me, MeO) are oxidized though an ET pathway. In conclusion, we have shown that both HAT and ET mechanisms occur for the oxidation of R-DMAs by dioxygen adducts 1R′. The reaction pathways are controlled by changes in the ease of substrate one-electron oxidation and also the reduction potentials of 1R′ (which are determined by ligand electronics).7,8 Coupled to all of this will be changes in the pKa’s of the bis-µ-oxo-ligands in 1R′, with stronger donor ligands (R′ ) Me2N and MeO) expected to produce better oxo bases (as H+ acceptors). Further investigations are needed to sort out these details. Acknowledgment. This work was supported by the NIH (K.D.K., GM28962; J.S., GM 067447). Supporting Information Available: Experimental details, KIE profiles, and linear free-energy plots (PDF). This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Solomon, E. I.; Sundaram, U. M.; Machonkin, T. E. Chem. ReV. 1996, 96, 2563-2605. (2) (a) Que, L., Jr.; Tolman, W. B. Angew. Chem., Int. Ed. 2002, 41, 1114. (b) Stack, T. D. P. Dalton Trans. 2003, 1881-1889. (3) Francisco, W. A.; Blackburn, N. J.; Klinman, J. P. Biochemistry 2003, 42, 1813-1819. (4) Pidcock, E.; Obias, H. V.; Zhang, C. X.; Karlin, K. D.; Solomon, E. I. J. Am. Chem. Soc. 1998, 120, 7841-7847. (5) Itoh, S.; Taki, M.; Nakao, H.; Holland, P. L.; Tolman, W. B.; Que, L., Jr.; Fukuzumi, S. Angew. Chem., Int. Ed. 2000, 39, 398-400. (6) Mahapatra, S.; Halfen, J. A.; Tolman, W. B. J. Am. Chem. Soc. 1996, 118, 11575-11586. (7) Zhang, C. X.; Liang, H.-C.; Kim, E.-i.; Shearer, J.; Helton, M. E.; Kim, E.; Kaderli, S.; Incarvito, C. D.; Zuberbu¨hler, A. D.; Rheingold, A. L.; Karlin, K. D. J. Am Chem. Soc. 2003, 125, 634. (8) Henson, M. J.; Vance, M. A.; Zhang, C. X.; Liang, H.-C.; Karlin, K. D.; Solomon, E. I. J. Am Chem. Soc. 2003, 125, 5186-5192. (9) The R′ group identity influences the peroxo/bis-µ-oxo equilibrium (more bis-µ-oxo isomer is present with more electron-donating R′)8 and rates of THF and MeOH oxidation (>1500× increase in THF oxidation rate for R′ ) Me2N vs R′ ) H).7 (10) (a) Karki, S. B.; Dinnocenzo, J. P.; Jones, J. P.; Korzekwa, K. R. J. Am. Chem. Soc. 1995, 117, 3657-3664. (b) Goto, Y.; Watanabe, Y.; Fukuzumi, S.; Jones, J. P.; Dinnocenzo, J. P. J. Am. Chem. Soc. 1998, 120, 1076210763. (c) Dinnocenzo, J. P.; Karki, S. B.; Jones, J. P. J. Am. Chem. Soc. 1993, 115, 7111-7116. (d) Manchester, J. I.; Dinnocenzo, J. P.; Higgins, L. A.; Jones, J. P. J. Am. Chem. Soc. 1997, 119, 5069-5070. (e) Baciocchi, E.; Lanzalunga, O.; Lapi, A.; Manduchi, L. J. Am. Chem. Soc. 1998, 120, 5783-5787. (f) Okazaki, O.; Guengerich, F. P. J. Biol. Chem. 1993, 268, 1546-1552. (g) Guengerich, F. P.; Yun, C.-H.; Macdonald, T. L. J. Biol. Chem. 1996, 271, 27321-27329. (11) (a) Sato, H.; Guengerich, F. P. J. Am. Chem. Soc. 2000, 122, 80998100. (b) Stearns, R. A.; de Monetellano, P. R. O. J. Am. Chem. Soc. 1985, 107, 4081-4082. (c) Hall, L. R.; Hanzlik, R. P. J. Biol. Chem. 1990, 265, 12349-12355. (12) Yields for R-MA are 1H, ∼60%; 1MeO, ∼80%; and 1Me2N, ∼90%. (13) See Supporting Information. (14) Yields are low, and 18-O incorporation in benzaldehyde varied from 36% to 68%. We suspect the low yields are due to unfavorable steric interactions between the dibenzyl groups and the Cu2(O2) core, and that the low isotope incorporation is due to exchange of the carbonyl oxygen with residual water in the solvent.13 (15) Johnson, C. D. The Hammett Equation; Cambridge University Press: Cambridge, U.K., 1973. (16) Going from a 10- to 100-fold excess of substrate did not change the relative yields. (17) All KIE values are within the semiclassical limit for ET and HAT reactions at -80 °C (kH/kD could reach a maximum of 23.1). (18) Shearer, J.; Karlin, K. D., unpublished results. For example, 1H will oxidize certain ferrocene derivatives that 1Me2N will not.
JA0359409 J. AM. CHEM. SOC.
9
VOL. 125, NO. 42, 2003 12671
